In migrating from Hibernate Search 3.1.1 to 3.2CR1, the master server will just do a loop with all JMS messages if the backend, connection_factory, queue and backend are configured as with a slave indexing server. It reads a JMS message with indexing operations and then sends a JMS indexing operation message right back.

So my question is how should a master indexing server write indexes directly? Or can it? I want to be able to tell the indexing server to index a specified number of entities directly. If I don't configure any of the above properties, will it properly write the indexes?

I'm sorry I don't think I understand the question. Have you read the reference documentation of the master slave configuration? Or Hibernate Search in Action? The master index process is soley responsible for updating the indexed. Slave nodes make requests for updates via JMS.



What do you mean by this?

Basically I want to use my master indexing server for 2 purposes:

Read from the Hibernate Search JMS queue to write the master index
Manually index specified entities when I specify to do so over the web interface

Is this possible? What's the proper settings for the connection_factory, queue, execution and backend?